Installed packages

DuckDB runtimeDownload PDF (2.8 MB)

Packages without a version install at their latest release when the runtime is built. To read the exact version in your session, run pip show <name> in a notebook.

Python and system#

Component	Version or detail
Python	3.12
`numpy`	2.3.5
`pandas`	2.2.0 or newer, below 3
`pyarrow`, `SQLAlchemy`, `boto3`, `urllib3`, `matplotlib`, `pymysql`	unpinned
`jupyterlab`, `ipython`, `ipykernel`	unpinned
`mlflow`	2.20.4
`torch`, `torchvision` (CPU)	2.14.0, 0.29.0
`scipy`	1.17.1
`scikit-learn`	1.5.2
System tools	AWS CLI v2, `unixodbc`, PostgreSQL and MySQL client libraries, time-zone data

Time zone data is included, so time-zone-aware timestamps and the UTC zone work correctly in PyArrow.

Connector and data libraries#

Package	Version	Purpose
`duckdb`	1.5.5	Query engine
`qualipilot[gz,linking]`	3.5.0	Data quality and record linkage
`pymongoarrow`, `pymongo`	unpinned	MongoDB
`google-cloud-bigquery-storage`, `google-cloud-bigquery`	unpinned	BigQuery
`pyodbc`	unpinned	ODBC databases
`psycopg2-binary`, `cx-Oracle`	unpinned	PostgreSQL, Oracle
`teradatasql`	unpinned	Teradata
`snowflake-connector-python`, `snowflake-sqlalchemy`	unpinned	Snowflake
`s3fs`	unpinned	S3 file system
`simple-salesforce`	unpinned	Salesforce
`azure-storage-blob`	unpinned	Azure Blob Storage
`google-cloud-storage`, `gcsfs`	unpinned	Google Cloud Storage
`paramiko`	unpinned	SFTP
`cassandra-driver`	unpinned	Cassandra and ScyllaDB
`redshift-connector`	unpinned	Amazon Redshift
`requests`, `httpx` (>=0.27,<1), `httpcore` (>=1,<2)		HTTP clients
`azure-cosmos`	unpinned	Cosmos DB
`elasticsearch`, `opensearch-py`	unpinned	Elasticsearch, OpenSearch
`pyexasol`	unpinned	Exasol
`neo4j`, `graphdatascience`	unpinned	Neo4j
`pinecone`	unpinned	Pinecone
`weaviate-client`	unpinned	Weaviate
`kdbai-client`	>=1.3.0	KDB.AI
`firebolt-sdk`	unpinned	Firebolt
`influxdb3-python`	unpinned	InfluxDB 3
`ms-dynamics-business-central-sdk`	unpinned	Business Central
`shopifyapi`, `marketorestpython`	unpinned	Shopify, Marketo
`azure-synapse`	unpinned	Azure Synapse
`google-api-python-client`, `google-auth`, `google-auth-httplib2`, `google-auth-oauthlib`	unpinned	Google Drive and authentication
`msal`	unpinned	Microsoft identity
`couchbase`	unpinned	Couchbase
`clickhouse-connect`	unpinned	ClickHouse
`databricks-sql-connector`	unpinned	Databricks SQL
`splunklib`, `splunk-sdk`	unpinned	Splunk
`firebase-admin`	unpinned	Firebase and Firestore
`adbc-driver-postgresql`, `adbc-driver-sqlite`, `adbc-driver-manager`	unpinned	Arrow database drivers
`pyiceberg[sql-postgres,s3fs,pyarrow]`	unpinned	Iceberg client
`hudi`	unpinned	Apache Hudi
`pyathena`	unpinned	Amazon Athena
`kafka-python`	unpinned	Kafka
`azure-eventhub`	unpinned	Event Hubs
`singlestoredb`	unpinned	SingleStore
`openpyxl`	unpinned	Excel files

LLM and AI libraries#

Package	Version	Purpose
`openai`	unpinned	OpenAI and Azure OpenAI
`anthropic`	unpinned	Anthropic
`cohere`	unpinned	Cohere
`huggingface-hub`	unpinned	Hugging Face
`google-generativeai`	unpinned	Gemini
`mcp`	>=1.29,<2	Model Context Protocol client

Document extraction#

Package	Version	Purpose
`docling[rapidocr]`	2.116.0	Document conversion with OCR
`docling-core`	2.88.0	Docling core
`docling-ibm-models`	3.13.3	Docling models
`docling-parse`	7.8.1	Docling PDF parser
`rapidocr`	3.9.2	OCR
`onnxruntime`	1.28.0	Model inference
`pdf2image`	>=1.17	PDF rasterizing
`pdfplumber`	>=0.11	PDF text extraction
`pillow`	>=10.0	Images
`pypdf`	>=4.2	PDF utilities
`pytesseract`	>=0.3	OCR bridge

The layout, table-structure and OCR models come with the runtime, so PDF extraction works without internet access. The Groundzero helper packages (GZ and others) are already importable.

DuckDB extensions#

These extensions are installed. Connectors load the ones they need the first time you use them. For the others, such as spatial, run session.query("LOAD spatial"). Nothing is downloaded when your code runs.

Extension	Used for
`httpfs`	Reading and writing S3, GCS and HTTP objects
`aws`	AWS credential handling
`iceberg`	Groundzero Lakehouse and Iceberg tables
`parquet`, `json`	File formats
`azure`	Azure Blob Storage
`postgres`, `mysql`	Attach and write PostgreSQL and MySQL databases
`spatial`	Geospatial types and functions
`motherduck`	MotherDuck

Packages added by the ML variants#

Variant	Packages
scikit-learn	`opencv-python-headless`, `xgboost`, `lightgbm`, `catboost`
PyTorch	`torch` 2.5.1, `torchvision` 0.20.1 (CUDA 12.1), `opencv-python-headless`, `xgboost`, `lightgbm`, `catboost`, `shap`, `lime`, `albumentations`, `huggingface_hub`, `transformers`, `simsimd`, `fastai`
TensorFlow	`tensorflow`, `opencv`, `xgboost`, `lightgbm`, `catboost`, `shap`, `lime`, `nltk`, `spacy`, `h5py`